We analyze theoretically the existence of precursor anomaly migration and the nonlinear variation of migration velocity.

Synergistic effects, nonlinear variation, and difficulties in discounting the future make the intellectual problem difficult, but not (in principle) insoluble.

By introducing the nonlinear variation weight and mutational operation into the standard particle swarm algorithm to ensuring the overall convergence and enhance the accuracy of convergence.